Mendeley

Mendeley

Mendeley is a reference management system for scholarly work. It stores citation metadata separately from uploaded full-text files, organizes records into personal or shared libraries, and supports collaboration through private or invite-only groups.

Data Model

• Profile — researcher identity, institution, academic status, and disciplines.
• Catalog records — global bibliographic entries useful for discovery and citation import. Catalog results are not automatically part of a user's library.
• Documents — library citation records. A document can be personal or group-scoped, can have metadata such as authors, source, identifiers, tags, notes, read/starred state, and can exist without an attached file.
• Files — binary attachments, usually PDFs, linked to documents. File deletion removes the attachment, not the citation record.
• Folders — personal or group organization containers. Removing a document from a folder does not delete the document. Deleting a folder does not delete its documents.
• Groups — shared research libraries. Current group creation should use private or invite-only access; public group creation is blocked by Mendeley.
• Annotations — reading notes, highlights, and marked regions attached to documents or files.
• Trash — soft-deleted documents. Restore returns the citation record, but original folder membership is not restored.

Common Workflows

• Discover a paper in the global catalog by DOI, identifier, title, author, or topic; then create or update a library document from the returned citation metadata.
• Search personal or group library documents when the paper should already be in the account. Use catalog discovery when the paper may not be in the library yet.
• Create a document before uploading a PDF; files attach to existing documents rather than replacing document metadata.
• Organize documents by creating folders, adding document IDs to folders, and removing folder membership when moving between organizational structures.
• Work in shared libraries by selecting a group and using that group context for documents, folders, annotations, and files.
• Use soft deletion first for documents that may need recovery. Permanent deletion belongs only after trash review.
• Folder downloads: when the user wants PDFs from a folder, use the folder batch-download operation with cursor pagination (has_more, next_cursor) until the folder is complete — or list folder attachments first to inspect sizes. Do not loop per-document file metadata once per library record. Small PDFs return content_base64; larger ones return presigned download_url with spillover: true.

Operational Notes

• Document IDs, file IDs, folder IDs, annotation IDs, and group IDs identify different object types and are not interchangeable.
• Listing attachments for a document returns metadata (size, mime type, hash) — never PDF bytes. Use single-file download or folder batch download for binary content.
• File download returns base64 content and metadata. Large outputs may be delivered as a resource link by the platform.
• PDF uploads require a target document, filename, MIME type, and base64 content.
• Group deletion removes a collaboration space. Folder deletion only removes an organization container. File deletion removes an attachment. Document deletion affects the citation record.
• Catalog search and library search can hit upstream rate limits; narrower identifier lookups or smaller result limits are more reliable than broad keyword searches.